arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2007.04309cs.LGcs.CVcs.ROstat.ML
收藏

部署过程中的自我监督策略自适应

Self-Supervised Policy Adaptation during Deployment

IRI, CSIC-UPC(加泰罗尼亚理工大学西班牙国家研究委员会机器人研究所);UC Berkeley(加州大学伯克利分校);NYU(纽约大学);UC San Diego(加州大学圣迭戈分校);Technical University of Denmark(丹麦技术大学)

Nicklas Hansen, Rishabh Jangir, Yu Sun, Guillem Alenyà, Pieter Abbeel, Alexei A. Efros, Lerrel Pinto, Xiaolong Wang

摘要:

在大多数真实世界场景中,在一个环境中通过强化学习训练的策略需要部署在另一个可能完全不同的环境中。然而,在不同的环境中推广是困难的。一个自然的解决方案是在新环境中部署后继续训练,但如果新环境没有提供奖励信号,这就无法做到。我们的工作探索了自我监督的使用,允许政策在部署后继续培训,而不使用任何奖励。虽然以前的方法明确地预测在新的环境中的变化,我们假设没有这些变化的先验知识,但仍然获得显着的改进。经验评估是在DeepMind Control套件和ViZDoom的各种模拟环境中进行的,以及在不断变化的环境中进行的真实机器人操作任务,从未校准的相机中进行观察。我们的方法在36个环境中的31个环境中提高了各种任务的泛化能力,并在大多数环境中优于域随机化。

英文摘要:

In most real world scenarios, a policy trained by reinforcement learning in one environment needs to be deployed in another, potentially quite different environment. However, generalization across different environments is known to be hard. A natural solution would be to keep training after deployment in the new environment, but this cannot be done if the new environment offers no reward signal. Our work explores the use of self-supervision to allow the policy to continue training after deployment without using any rewards. While previous methods explicitly anticipate changes in the new environment, we assume no prior knowledge of those changes yet still obtain significant improvements. Empirical evaluations are performed on diverse simulation environments from DeepMind Control suite and ViZDoom, as well as real robotic manipulation tasks in continuously changing environments, taking observations from an uncalibrated camera. Our method improves generalization in 31 out of 36 environments across various tasks and outperforms domain randomization on a majority of environments.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑